56 minute read
FAD On-Site Visits
the counties from their pollbook, ballot activity report uploads to the statewide database (TEAM), ballot by mail rosters, final vote history, and canvass reports. The counties complied with the requests in varying degrees.
In reviewing the documents provided by the counties, it is noteworthy that each county records their election data differently and the counties use different vendors, which leads to differences in the data captured and reports available. Additionally, each county’s SVC or EVBB has developed its own processes, procedures, or forms to handle mail ballots, resulting in different points in each of the counties at which data could be evaluated and analyzed.
County Procedures
Considering every county has different procedures, forms, and practices, it was important for FAD to speak with members of the county’s staff that handled BBMs, members of the EVBB, and members of the SVC if possible to understand those processes and how to analyze the data provided by the counties. Moreover, Dallas and Harris Counties experienced changes in leadership from the time of the 2020 General Election to the time of the audit, making information from those with the best knowledge regarding practices and procedures during the 2020 General Election critical.
During on-site visits to Collin, Dallas, and Tarrant Counties, leadership, staff, and members of the EVBB/SVC were available to FAD for questions regarding processes and procedures. These three counties also provided FAD with tours of the facilities that showed the mechanics of EVBB/SVC operations. During one of the on-site visits to Harris County, the serving Elections Administrator and her Deputy made themselves available for general questions regarding the 2020 General Election. After that trip, Harris County emailed a list of certain members of the SVC, EVBB, and Central Count team that FAD could contact. Other than the Elections Administrator and her Deputy, Harris County staff was not identified nor made available for questions regarding procedures for voting by mail in the 2020 General Election by Harris County.
Across all of the counties, a consistent theme emerged. There were great difficulties in processing and handling the unprecedented volume of mail ballots issued during the 2020 General Election.
Collin County
In 2020, Collin County did not have an SVC, instead operating only with an EVBB. When a signed carrier envelope was returned to Collin County, the carrier envelope was scanned using VOTEC software. Two-person teams on the EVBB, with one member from each of the Democratic and Republican Party, reviewed electronic images of the voter’s BBM and returned carrier envelope. If the team did not agree that the signatures matched, the review would be forwarded to the Ballot Board Judge and Alternate Judge. A consensus of the EVBB would then vote on whether to accept or reject the ballot. If the ballot was accepted, the EVBB opened the carrier envelope and prepared the ballot for tabulation.
Collin County’s EVBB did not wait for ballot reading problems and instead was proactive in reviewing BBMs for issues that might cause problems when the ballot was scanned. Collin County developed a form to document and track the ballots being duplicated and the numbers associated with both the original and duplicated ballot. This was unique to Collin County and is a best practice. The ballots that had been prepared for tabulation were packaged and sealed for transfer to the counting station. These ballots were accompanied by a transmittal sheet that recorded how many ballots were being transferred and contained the signatures of the Ballot Board Judge and Alternate Judge.
The Ballot Board Judge also served in the counting station. Upon arrival at the counting station, the seal numbers were verified and if the seal numbers matched, the ballots were scanned for tabulation. In the event a ballot was rejected, the reason for the rejection was documented and kept separate so it would not be counted. The Ballot Board Judge ensured that the voter was mailed a letter notifying them that their ballot had been rejected and the reason for such.
Dallas County
Dallas County experienced difficulties in processing the high volume of ballots by mail due to staff turnover that occurred just prior to the 2020 General Election. The turnover lead to an influx of temporary workers and an inability to effectively train the temporary workers. In addition, although Dallas County purchased equipment to process ballots more quickly, it did not arrive in a timely manner and was ultimately unreliable. Dallas County also experienced issues with the USPS not delivering BBMs to voters in time, so Dallas County developed an internal procedure with dates by which to send voters their ballots in order to ensure them being timely received.
Dallas County separated ABBMs into categories such as yearly, regular mail, disabled and military applications. The applications were logged into Dallas County’s voter
registration system (VEMACS). Applications for voters who did not qualify for a BBM were not logged and were physically stored in a box labeled “Bad Mail.” If there was an issue with an ABBM, Dallas County mailed the voter a letter with a second ABBM. If the voter was not registered, Dallas County mailed the voter a letter with an application for registration and a second ABBM.356
Dallas County’s processing system involved printing labels for the green outer envelope mailed to the voter, the brown carrier envelope for the voter to return to Dallas County, and the jacket envelope. These labels contained the voter’s name, address, coded reason for voting by mail, certificate number, envelope identification number, and the election code. USPS requested that Dallas County put the return address mailing label on the back of the brown envelope. Unfortunately, this placement resulted in voted ballots being returned to the voter instead of being sent to the county. Dallas County corrected this issue by blacking out the addresses to ensure ballots mailed by voters were actually returned to the elections department.357
When BBMs arrived in Dallas County, the carrier envelopes were date stamped and run through a machine that imaged the envelope. Staff printed the voter’s ABBM for later comparison with the signature on the BBM. The carrier envelopes were batched and set aside for pick up and review by the SVC.
Members of Dallas County’s EVBB and SVC simultaneously served in both capacities.358 The batches of carrier envelopes and their corresponding applications were reviewed by two-member teams from the SVC.359 Each two-member team was comprised of one Democratic and one Republican member. The two-member team reviewed the carrier envelope and ABBM for signature comparison. Training included examples of what to look for and how to evaluate the signatures on the envelopes and applications. If the two-member team agreed that both signatures belonged to the voter, the carrier envelope moved on to be batched for opening and further
356 Dallas County indicated that now, they do not send an application for a ballot by mail to a voter that is not registered. Instead, they send a voter registration application. Dallas County also attempts to verify whether the voter is eligible to vote using a limited ballot. 357 Dallas County said they have since changed their process, utilizing labels now for the green and carrier envelopes that do not print the address of the voter on them. 358 The SOS recommends members of the Ballot Board be different than those that comprise the Signature Verification Committee. See The Office of the Secretary of State, Elections Division, Ballot Board Handbook 3, 2022 available at: https://www.sos.state.tx.us/elections/forms/ballot-boardhandbook.pdf#search=early%20voting%20ballot%20 board (“IT is NOT RECOMMENDED that members who serve on the EVBB also serve on the SVC. This creates a conflict because if a SVC has determined that the signatures on the application for ballot by mail or carrier envelope are not those of the same person, the EVBB may make a determination that the signatures are those of the same person by a majority vote of the board’s membership.”). Dallas County has since changed its procedures and the members of the SVC and EVBB do not serve simultaneously in both capacities. 359 This process in Dallas County is now electronic and does not involve the hand transfer of ballots.
processing. If the two-member team did not agree that the signatures matched, a second two-member team was asked to review the carrier envelope and application. If that team did not agree, the chair of the SVC determined whether to send it to the EVBB for further review.
The EVBB met in committees of between two and four members, with the judge serving as the third and fifth member of the committee. The EVBB would review and vote on questioned carrier envelopes. If the vote resulted in a tie, the judge served as the tie-breaking vote.
Carrier envelopes that were accepted were batched and processed by a team of two individuals. This team opened the carrier envelope and separated the internal secrecy envelope containing the ballot from the external carrier envelope. The team continued on with this process until they had a number of ballots to transfer to Central Count. These ballots were then grouped and sealed for transfer. A transmittal sheet documenting the number of ballots to be transferred and the seal number accompanied the ballots transferred to Central Count.
Upon the initial scan of the carrier envelope, the ballots were coded with a return status that indicated they were pending review. The date the scan took place was captured. After the carrier envelopes had been reviewed and a decision was made regarding whether to move them forward in the process, the carrier envelopes were scanned again and the status of the ballot changed from pending review to “OK.” The system also captured this second scan date. In reviewing electronic records from Dallas County, it appeared many ballots were returned past the deadline, yet had been counted. Dallas County explained the scanning process and it appears the scan date overrode the date of the ballot’s return to Dallas County.
Harris County
Members of Harris County’s SVC & EVBB served on the same board/committee. Harris County utilized a tub system to transport and process BBMs. The BBM process was described to FAD as “chaotic.” One of the reasons attributed to this description was the volume of ballots by mail. The SVC/EVBB involvement in the BBM process all occurred at the NRG Arena.
On receipt, returned carrier envelopes were scanned in batches of 25. These batches were placed in tubs that were transferred to the EVBB. Tubs were handled by twoperson teams of a representative from the Democratic and Republican parties. On arrival, the tub was opened and the carrier envelopes were counted to verify that the number of envelopes to be reviewed matched the number of envelopes that had been transferred. Using digital images of the carrier envelopes and applications, the teams
performed signature comparison. Any questionable carrier envelope was removed from a batch and wrapped in a separate sheet for further review. The original batch was re-wrapped or sealed with the batch sheet around it with an updated number to reflect any questionable carrier envelopes that had been removed. When a tub was completed, all of the carrier envelopes that were deemed “OK” were sealed back into a tub. The questioned carrier envelopes were kept separately for further review and packaged in a judge’s tub.
A three-tier system was used to evaluate questioned carrier envelopes. If the firsttier team could not agree regarding signatures, the carrier envelope was “questioned.” The carrier envelope would proceed to a second tier or more experienced team. Usually at this stage, many questioned carrier envelopes were resolved. In the event the second-tier team could not resolve the matter, the carrier envelope went to the judges for review. If there was nothing in the records available to the judges to verify that the signatures could belong to the same voter, the carrier envelope was rejected. A rejected carrier envelope was returned to the elections department and a voter was mailed a letter and reason for the ballot’s rejection.
The sealed carrier envelopes deemed “OK” were returned to the elections department. At this stage, the carrier envelopes were opened to process the actual ballots. At times, the EVBB helped with this process if they had availability. First, the carrier envelope was opened and the internal secrecy envelope containing the ballot was removed. The carrier envelope was torn halfway down through the voter’s address. The secrecy envelopes were stacked. If a ballot arrived in a carrier envelope that was not sealed in a secrecy envelope, the team member called for a secrecy envelope, placed the ballot inside the secrecy envelope, and put it inside the envelope.
The ballot envelopes were then opened and any ballots that had questionable marks making the intent of the voter a question had to be adjudicated. Additionally, if a voter had marked through or on the barcode, the ballot had to be duplicated.
Tarrant County
Members of the EVBB and SVC served on the same board/committee.360 While onsite in Tarrant County, FAD had the opportunity to observe BBM processing for a subsequent election. Tarrant County developed forms for tracking the movement of ballots between stations to ensure accurate accounting of the ballots. Tarrant County noted that ballots by mail are not tracked until a voter is sent a ballot, but that a
360 This set up is contrary to best practice.
future project will likely involve digitizing records of all the applications that Tarrant County receives to enable tracking of the applications received.
Signed returned carrier envelopes were checked in by running the envelopes through a sorting machine that captured an image of both the front and back of the carrier envelope and scanned the barcode to note the envelope had arrived. Signature comparison between the signed carrier envelopes and the voter’s application to vote by mail was done by two-member teams consisting of one Republican and one Democratic member. This process was done electronically utilizing large monitors. Once the signature comparisons were completed, the envelopes were run back through the sorter and the carrier envelopes that had been accepted would be cut open by the machine. The EVBB extracted the ballots, put them in batches, prepared them for scanning, and scanned them using the Hart Verity Scan.361 On the day of tabulation, the ballots were placed on a vDrive and given to Central Count to add to the system for tabulation. This physical transfer of the electronic information storage media was done without a log to document the transfer. FAD made Tarrant County aware that this transfer requires documentation and Tarrant County intends to address it.
If the two-member team could not agree regarding the signature comparison, the EVBB would hold a session to look at the questioned carrier envelopes and vote as a group on whether to accept or reject it. If there was a tie, the judge made the final decision. Tarrant County’s early voting mail staff assisted the EVBB by generating letters regarding rejected carrier envelopes and provided them to the EVBB for signature and mailing.
Tarrant County experienced an issue with the print quality of their BBMs that required the ballot board to duplicate the ballots following the procedure set out in the Election Code.362 Tarrant County stamped both the initial ballot and the duplicated ballots with a code so that they could be matched and stored the physical ballot pairs in 25 boxes. They subsequently scanned the ballots from the 25 boxes and stored them along with their tracking logs in a set of 56 Adobe Acrobat files. Given the large volume of data, FAD took a sample of ballots from each box and reviewed a total of 160 ballot pairs. Of these pairs, FAD located two duplication errors.
361 The extraction and scanning process occurs in the same room. 362 See Duplicated Ballots, supra/infra.
Figure 8-3: Original Ballot 9000312
Figure 8-4: Duplicate Ballot 90000312
Figure 8-5: Original Ballot 30001815
Figure 8-6: Duplicate Ballot 3001815
The vendor had provided a test batch that worked, but the ballots that were ultimately printed were rejected by the scanners. Tarrant County switched vendors and added protections in its contract with the new vendor to enhance testing and reduce the odds of this reoccurring. This issue has not reoccurred since.
BBM Data Analysis
FAD endeavored to determine, based on the information provided by the counties:
• The number of applications for ballot by mail that a county received;
• The number of applications that were approved;
• The number of applications that were rejected;
• The number of ballots by mail that were returned;
• The number of ballots by mail that were accepted;
• The number of ballots by mail that were rejected; and
• The number of ballots by mail ultimately tabulated as reported in the canvass.
In addition to reviewing the electronically available documents provided by the counties, FAD engaged in multiple on-site visits at each county’s election department to review, inspect, and obtain copies of election records related to ballots by mail. The records available and the usefulness of the records depended largely on the county’s organization of and inventory of said records. FAD also met with individual election staff members, signature verification committee members, or early voting ballot board members to gain insight into the processes and procedures in place for the respective counties.363 Changes in leadership in Dallas and Harris Counties that occurred after the 2020 General Election, highlighted the importance of speaking to staff or members of the SVC or EVBB who would have the best understanding of each county’s processes and procedures as they existed during the 2020 General Election.
Applications for Ballots by Mail
In 2020 none of the four counties had a mechanism in place to log or track the mere receipt of an ABBM until the application was accepted. Unless and until a determination was made that a ballot would be mailed to the prospective voter, the ABBM was not logged or tracked in any manner. If an ABBM came in and that application was rejected, i.e., no ballot would be mailed to the voter, there was no log or database report available capturing this occurrence. Collin and Tarrant Counties, however, were able to provide scanned images of all applications received by the county.
Due to the lack of recording or tracking of all ABBMS (regardless of whether a ballot was ultimately mailed to a voter) received by the counties, meaningful analysis of the figures the county provided in response to the Secretary of State request for the total number of ABBM received was not possible. Further, as the counties did not begin tracking the application until a ballot was mailed, the counties generally indicated there was no real mechanism by which to evaluate whether an ABBM was rejected.364 Using the data available, FAD was able to determine the numbers that follow regarding how many applications resulted in ballots being sent to a prospective voter. The values available from the documentation provided by the counties is listed below in comparison to the written response as reported in the initial letter to the
363 Prior to Harris county’s recent cooperation beginning in October 2022, Harris County did not make staff available for comment or information regarding processes or procedures during the 2020 General Election. Harris County provided contact information for certain individual members of the EVBB, SVC, or Central Count “we”. FAD spoke with each of the individuals that were identified by Harris County. Even so, a more thorough understanding of processes and procedures during the 2020 General Election would have been possible with input from Harris County staff members earlier in the auditing timeline. 364 Tarrant County, however, provided a list of the letters mailed to rejected applicants and the number of rejection notices sent regarding rejected applications for BBM was 3,180.
Secretary of State. Unique findings related to ABBMs in certain counties are included below.
County Letter Response Records Provided by County
Collin 43,814
Dallas 103,220 Absentee Request Type and Return Status Summary: 45,948365
1120 Summary Report for 2020 General Election: 103,220366
Election Audit Workbook: 101,409367
Harris 206,020 Ballot List–By Mail–All Return Statuses: 252,693368
Tarrant 89,305 Absentee List All Return Codes: 89,305369
Figure 8-7: Ballots by Mail Sent to Voters
Reason for Requesting Ballot by Mail
The two most common categories of BBM voters are individuals who vote by mail due to age and disability. Voters are not required to provide the specific nature of their disability on an ABBM but are required to certify that their disability prevents them from appearing at the polling place without the likelihood of needing personal assistance or injuring their health. Available records do not provide any means for auditing a request to vote by mail due to disability. Available records did, however, provide means for auditing requests to vote by mail due to age.
In order to have been eligible to vote by mail due to age in the 2020 General Election, a voter must have been born on or before November 3, 1955. FAD evaluated records
365 This document breaks down the categories of requests of BBM type, i.e. Annual Disability, Overseas Military, Expected Absence from County, etc. with a cumulative total of all requests 366 This document is a report of all absentee labels printed with a breakdown of those accepted, rejected, and not returned 367 This figure comes from Dallas’ in-house audit workbook of ballots issued vs. ballots counted 368 This is a cumulative report of all voters that requested a ballot 369 This document breaks down the categories of requests of BBM type, i.e. Annual Disability, Overseas Military, Expected Absence from County, etc. with a cumulative total of all requests
related to ABBMs and BBMs to verify that voters coded as voting by mail due to age were in fact 65 or older at the time of the 2020 General Election.
Two source documents provided by Collin County were used for this portion of the audit: Collin County’s Final Vote History and Collin County’s Absentee Returns roster dated November 4, 2020.
Collin County’s Final Vote History reflected there were 84 voters coded as yearly applicants to vote by mail for age with dates of birth indicating they were younger than 65.
Collin County’s Absentee Returns roster from November 4, 2020 reflected there were 198 voters coded as yearly applicants to vote by mail for age with dates of birth indicating they were younger than 65.370
In reviewing the applications to determine whether there was a coding error or these individuals improperly received ballots by mail, it was determined that:
Of the 85 voters in Collin County’s Final Vote History371
• 77 voters had actually requested to vote by mail for disability, 6 voters had requested to vote by mail due to an expected absence from the county, and 1 voter requested a ballot due to being overseas. These 84 voters were coded incorrectly and were entitled to vote by mail.
• One voter did not include a reason for voting by mail on their ABBM and did not include a date of birth. This voter was not entitled to vote by mail.
Of the 198 voters in Collin County’s Absentee Returns roster from November 4, 2020:
• 168 voters had actually requested to vote by mail for disability and 6 voters had requested to vote by mail due to an expected absence from the county.
These 174 voters were coded incorrectly and were entitled to vote by mail.
• 2 voters had the incorrect date of birth associated with them and were actually 65 or older. These voters were entitled to vote by mail.
370 There were 8 voters in common between the two source documents. 371 Collin County’s Final Vote History file was incomplete in that it did not reflect the reason for voting by mail on 25,577 voter records.
• 2 voters had improperly been coded as being associated with a BBM. In each of these instances, the ballots for these voter records had been requested by another voter with a similar name who was actually 65 or older.
• 20 voters had applications that either reflected no reason for voting by mail or did in fact request to vote by mail due to being 65 or older, were issued a ballot, yet were not entitled to vote by mail.
Dallas County’s ballot by mail records indicated 994 voters were coded as yearly applicants to vote by mail for age with dates of birth indicating they were younger than 65. Dallas County pulled specific applications in response to FAD’s request to review these voters’ applications to vote by mail.
Of the 994 voters in Dallas County’s records:
• 775 voters had requested to vote by mail due to disability and 66 voters had requested to vote by mail due to an expected absence from the county. These 841 voters were coded incorrectly and were entitled to vote by mail.
• 59 applications were not found and no determination of eligibility could be made.
• 2 voters listed COVID as their eligibility reason and were improperly issued a ballot and were not entitled to vote by mail.372
• 22 voters requested to vote by mail but their applications were sent to an ineligible recipient.
• 2 voters listed both age and disability and were properly issued a ballot and entitled to vote by mail.
• 21 voters’ applications reflected no reason for voting by mail and were improperly issued a ballot and were not entitled to vote by mail.
• 42 voters had applications that requested to vote by mail due to being 65 or older but date of birth records indicated the voters were not 65. These voters were improperly issued a ballot and were not entitled to vote by mail.
372 See, In re State, 602 S.W.3d 549, 550, (2020) (“We agree with the State that a voter's lack of immunity to COVID-19, without more, is not a ‘disability’ as defined by the Election Code.”).
• 5 applications were illegible and no determination of eligibility could be made.
The records related to ballots by mail provided by Harris County did not provide the level of detail necessary to enable an audit of this category of voters. Documentation available only broke down absentee ballots by type as follows:
35 - Regular app voter overseas
DME - Domestic Military Email
DMM - Domestic Military Mail
DSE - Domestic Spouse/Dep Mil. Email
DSM - Domestic Spouse/Dep Mil. Mail
EB - EMERGENCY BALLOT-ILLNESS/DEATH
FCE - Federal Overseas Civilian Email
OCE - Overseas Civilian Email
OCM - Overseas Civilian Mail
OIE - Fed Overseas Indefinitely Email
OIM - Fed Overseas Indefinitely Mail
OME - Overseas Military Email
OMM - Overseas Military Mail
OSM - Overseas Spouse/Dep Mil. Mail
RM - REGULAR MAIL
RS - Regular at diff residence address
Further documentation regarding the reason for voting by mail was not provided. RM includes the vast majority of the BBM.
Two source documents provided by Tarrant County were used for this portion of the audit: Tarrant County’s Final Vote History and Tarrant County’s Absentee List with all return codes.
Tarrant County’s Final Vote History reflected there were 128 voters coded as yearly applicants to vote by mail for age with dates of birth indicating they were younger than 65.
Tarrant County’s Absentee List with all return codes reflected there were 282 voters coded as yearly applicants to vote by mail for age with dates of birth indicating they were younger than 65. In addition, this spreadsheet reflected there were 71 voters coded as voting by mail (non-yearly applicants) due to age with dates of birth indicating they were younger than 65.
FAD provided a list of these voters to Tarrant County and requested their ABBM to review whether this was due to an error in coding or an error in issuing the ballot itself.
Tarrant County indicated they were unable to pull specific applications but that upon review of the list FAD provided, it appeared many of the applications had been coded incorrectly. Tarrant County provided all of their ABBMs for FAD to locate and review those ABBMs at issue. In total, Tarrant County provided over 99,000 ABBMs in an unsearchable format. A review of a 5% sample of the ABBMs at issue revealed that all of the questionable ABBMs were the result of miscoding and the voters who had requested to vote by mail were entitled to do so. The majority of these voters had requested to vote by mail due to disability or an expected absence from the county, were coded incorrectly, and were entitled to vote by mail. Two voters in the sample had improperly been coded as being associated with a BBM. In both of these instances, the ballots for these voter records had been requested by another voter with a similar name who was actually 65 or older.
ABBM Unique Findings
Dallas County
Dallas County staff indicated that while there was no tracking mechanism for ABBMs that were received but ultimately did not result in the issuance of a ballot, Dallas County did maintain those applications. The applications were categorized as “Bad Mail” and were preserved in a box following the election. A review of a sample of those records revealed the following as reasons for rejection of the application:
• Missed deadline;
• Expected absence invalid;373
• No signature;
• Unable to determine;
• Signature mismatch;
• Not registered;
• No election selected;
• Invalid forwarding address; and
• No name.
Bulk Applications for Ballots by Mail
Dallas County’s records inventory indicated there were boxes containing “Applications Received in Bundle” for ballots by mail. FAD reviewed these boxes. The boxes contained over 1,464 applications for ballots by mail submitted in 97 envelopes for the 2020 Primary and General Elections. Each envelope contained between 1 and 93 applications. Some of these applications were from Secretary of State, as many voters mailed applications for ballots by mail to the Secretary of State and these were subsequently re-routed to the appropriate counties.
373 If the applicant indicating they were requesting a ballot by mail due to an expected absence from the county, the dates of the absence from the county must cover the date of the election for which the applicant is seeking a ballot.
Of the non-Secretary of State routed bundles, many included the name of an assistant that helped the applicant complete the ABBM. Many of these bundled applications were associated with common assistants. In particular, three assistants were listed on 469 ABBMs. A single person was named as the assistant for 393 ABBMs. Notably, these three assistants were also associated with repeated ABBMs for the same voters. Many of the applicants assisted by these three assistants had residential addresses on their applications associated with assisted living facilities and apartment complexes. One address in particular — identified as an assisted living facility — was the source of 55 ABBMs submitted with the same individual named as the assistant on all 55 applications. Another address — identified as an affordable housing complex — was listed on 58 ABBMs submitted with the same individual named as the assistant that had been named on the 55 applications from the assisted living facility. There was also one application submitted with this same assistant listed on the application that requested a BBM for the voter with age, i.e. being 65 or older, as the reason for voting by mail — while the application reflected the voter was only 57 years old.
FAD attempted to locate the carrier envelopes associated with the voter. This was difficult, however, because this required individual review of each of the over 70,000 scanned carrier envelopes. In reviewing a sample of those carrier envelopes, FAD located carrier envelopes executed by 22 of the voters that had been assisted by the assistant who was listed on 393 ABBMs. Of those 22 carrier envelopes, 2 listed the same assistant who was listed on the ABBM.
As noted above, a voter is entitled to assistance in filling out their ABBM. There is currently no statutory limit to the number of applications on which an individual can act as an assistant. Additionally, there is nothing in the Texas Election Code that prevents a person from being compensated for assisting with an application for a ballot by mail. Contrasted with the legal requirements and prohibitions associated with the actual mail ballot, there are fewer legal restrictions and requirements that apply to assisting with an ABBM. The Office of the Attorney General has agreed to assist with providing additional information so this occurrence in Dallas County may be referred to the local authorities for investigation and potential prosecution.
BBM Returned to the County
FAD endeavored to determine how many ballots by mail were returned to each county, how many of those ballots were accepted, and how many of those ballots were rejected. Finally, FAD attempted to verify that the number of ballots accepted were actually tabulated and that number matched what was reported by the county in the canvass. As will be evident in the findings below, the four counties had multiple
sources of data regarding BBMs, yet comparison and analysis of these sources yielded inconsistencies and discrepancies.
Collin County
Ballots Returned Accepted Ballots Not Accepted
Ballots Tabulated Canvass
37,663374 34,265375 7 36,007376 35,650377
Figure 8-8: Collin County Returned BBM Discrepancies
As noted above, there were discrepancies between the number of ballots returned, accepted, and rejected based on the records provided by Collin County. Collin County believes discrepancies observed may be due to how their reporting software categorized provisional and limited ballots. Collin County has since changed how these numbers are reported and they all have their own reporting group in the software.
Dallas County
Dallas County provided four electronic data sources for ballot by mail statistics in 2020. None of the sources are consistent. The following chart shows the source and numbers reflected in the source provided by Dallas County.
374 Absentee Returns through November 4, 2020. 375 Final Voter History_Nov2020_VEMACSexport. This includes 5 voters who voted by mail using a Limited Ballot. 376 Collin County Audit Log. 377 Combined Summary Results Report, “Ballots Cast – Total” for the Mail category.
Source Ballots Returned Accepted Ballots Not Accepted
ABBM Report378 72,119379 60,546380 11,573381
Ballot by Mail Voter Roster
List of Voters sent Mail Ballot
Election Audit Workbook386 77,617
73,265
91,919387 76,838382
61,533384
76,839
539383
11,545385
15,080388
Figure 8-9: Dallas County Returned BBM Discrepancies
Dallas County’s Final Vote History reflected 74,890 voters voted by mail. FAD also reviewed the Electionware Audit Log that contains a record of all ballots ultimately tabulated. The audit log reflects 78,174 ballots by mail were tabulated, which does not match any of the source documents provided by Dallas County. Dallas County’s canvass report states there were 78,147 ballots by mail cast in 2020.
Ballots Reported at Different Stages in Dallas County’s BBM Process
FAD reviewed records related to the chain of custody for ballot transfers by the EVBB to Central Count for tabulation. These records were packets that included a cover sheet with the Ballot Transit Case number that listed the contents, how many ballots
378 1120 All ABBM Report for 2020 General Election. 379 This figure is inclusive of multiple types of ‘return’ for the ballot including: ballots returned by a voter for normal processing and tabulation, ballots returned undeliverable, and ballots cancelled by a voter to vote in person. 380 1120 All ABBM Report for 2020 General Election; 1120 Summary Report for 2020 General Election. Including all ballots with the return status code OK. 381 1120 All ABBM Report for 2020 General Election; 1120 Summary Report for 2020 General Election. Including all returned ballots that were not coded OK. 382 Ballot by Mail Voter Roster. Including all ballots with a disposition code OK. 383 Ballot by Mail Voter Roster. Including all ballots that did not have a disposition code OK and were not returned undeliverable. There were 240 ballots coded as undeliverable. 384 Including all ballots with a disposition code OK. 385 Including all ballots that did not have a disposition code OK and were returned undeliverable. There were 187 ballots coded as undeliverable. 386 Source: VEMACS ABBM Tracking Table. 387 This number includes ballots that were returned undeliverable or cancelled by a voter. 388 This figure includes rejected ballots as well as those that were cancelled by a voter or returned undeliverable. This source reflects 784 ballots were rejected.
were included in the case, and the serial numbers of the ballot transit case. The second page of the packet contained a form that documented any irregularities regarding the ballots contained in the transit case such as: two ballots in one ballot envelope, no ballot in the envelope, or ballots that required duplication. This sheet contained a detailed accounting of the number of ballots that were in the case, their handling, and how many were transferred to the Tabulation Supervisor.
Figure 8-10: Cover sheet of packet Figure 8-11: Second page in packet.
Once the ballots went to tabulation, they were run through a DS850, DS450 or DS200. A report was generated following the scanning of the ballots entitled a Ballot Statistics Summary Report. This report shows when the ballots were run through the machine, how many sheets were processed and how many ballots were cast.
Figure 8-11: Ballot Statistics Summary Report
Ballot Statistics Summary from Ballot Transit Case #1
After each Ballot Statistics Summary page, there was a detailed list of the batches that comprised the transit case and the batch numbers assigned to sets of the ballots. This batch number could be used to identify when the ballots were ultimately tabulated by reviewing the Audit Events Report generated by Electionware. Every batch for which FAD obtained documentation in the Ballot Transit Case packets was identified by a letter and number combination.
Figure 8-12: Batch Reports from Ballot Statistics Summary
Figure 8-13: Election Audit Events Report
As noted above, there were discrepancies between the number of ballots transferred by EVBB to Central Count for tabulation. For example, ballot transit case 1 (Figure 810) contained 644 ballots, however only 635 ballots were processed for irregularities and ultimately ran through the ballot scanner (shown in Figure 8-12). There was no documentation to explain the nine-ballot difference. The ballot transit case documentation revealed 76,991 ballots were transferred by the EVBB but only 76,890 were processed for irregularities and delivered to the Central Count Supervisor.
An additional irregularity that was present in the ballot transit case packets was a discrepancy between the number of transferred ballots processed for irregularities and the number of ballots cast according to the Ballot Statistics Summary Report.
The documents below exemplify this occurrence:
Figure 8-14 Ballot Transit Case #5 transferred 608 ballots.
Figure 8-15: 607 ballots were processed for irregularities and submitted to the Tabulation Supervisor.
Figure 8-16: Ballot Statistics Summary sheet reflects 595 ballots cast for Ballot Transit Case #5
ES&S confirmed that the number reflected in the “Total Ballots Cast” portion of the report reflected the number of ballots cast. The difference between the “Total Sheets Processed” and “Total Ballots Cast” numbers would occur if there was a multi-page ballot that had been scanned, resulting in multiple pages being scanned but only corresponding to one ballot.
76,991 ballots 76,890 ballots 72,925 ballots 72,549 ballots with complete documentation
78,174 total ballots 78,147 ballots
Figure 8-17: Process of Reviewing and Recording Early Voting
According to Dallas County’s records, 76,991 ballots that had been processed by the SVC/EVBB between 10/17/2020 and 11/9/2020 were transferred by the EVBB to Central Count for processing and tabulation.
Analysis of the Ballot Transit Case documentation revealed that of the 76,991 ballots transferred to Central Count, only 76,890 ballots were processed for irregularities and delivered to the Central Count Supervisor.
According to the Ballot Statistics Summary sheets included in the packets for the 135 Ballot Transit Cases, of the 76,991 ballots transferred by the EVBB to Central Count 72,925 ballots were cast.389
As Dallas County’s canvass reflected there were 78,147 ballots by mail cast, further analysis was done using the Audit Events Report generated by Electionware to account for the BBMs that were tabulated. In reviewing the audit log, there were slight discrepancies between the Batch Statistics Summary sheets regarding the number of ballots cast and the number of ballots tabulated per batch in the audit log.
The total number of mail ballots tabulated according to the Audit Events Report was 78,174. In the Audit Events Report generated by Electionware, there were a total of 1,977 batches of ballots for which FAD had complete documentation in the Ballot Transit Case Packets.390 Those 1,977 batches as tabulated accounted for 72,549 ballots. There were 165 additional batches in the Audit Events Report coded as “mail” that FAD did not have complete documentation for in the Ballot Transit Case packets. Those additional batches accounted for 5,625 ballots. Of those 165 additional batches, there are 97 batches that used a different naming convention,391 lacked any supporting documentation, and accounted for 2,172 mail ballots.
Rejected Ballots by Mail
Dallas County’s SVC and EVBB records regarding rejected BBM were scanned for analysis and review. Dallas County also provided scanned documents regarding rejected BBM. Dallas County’s SVC and EVBB records included a log of the voters whose ballots were rejected, the letters sent to the voters notifying them that their ballot had been rejected, and copies of the forms documenting ballots referred by the SVC to the EVBB, and tally forms used by the EVBB to vote on whether to accept or reject the ballots. Records reflected that the EVBB met on October 20, November 3, November 9, and November 16, 2020 to review ballots referred to them by the SVC.
The worksheet and tally records for ballots referred by the SVC contained a detailed accounting of who served as the Ballot Board Judge, the board members participating in the meeting, and the date, time, and place of the meeting. Each member of the board that participated in the voting process completed their own individual
389 Four of the packets did not contain the Ballot Statistics Summary sheet. Further, some of the packets were incomplete or appeared to be missing pages. 390 Again, some of the Ballot Transit cases did not contain the Ballot Statistics Summary sheets or contained incomplete records; therefore, FAD was unable to trace certain batches back to the ballots transferred by the EVBB with the records provided. 391 For example, a batch number in a Ballot Transit Case is typically identified with a letter and number: D0148. The 97 entries lack the letter and number identifier and are simply coded EV Mail.
worksheet and tally record. The worksheets provided spaces to account for the ballot under review, the vote, and ultimate decision.
Figure 8-18: Worksheet and Tally Records
Records indicated that the SVC referred 79 ballots to the EVBB for further review. After review, 76 of those ballots were rejected and 3 were accepted. A breakdown of the reasons for rejection for those 76 ballots is as follows:392
Reason for Rejection
Two Ballots Enclosed
Signature Not Match
RJ 1
2 20
No Signature
No Ballot 25 28
Accepted 3
0 5 10 15 20 25 30
392 The reasons included in this table are the same terms used in the EVBB forms.
Figure 8-19: Rejection Statistics
Dallas County’s Early Voting Ballot Board Transmittal Form reflects the same total number of rejected ballots, but a slightly different breakdown regarding disposition.
Figure 8-20: Dallas County Early Voting Ballot Board Transmittal Form (Rejections)
Harris County
Ballots Returned Accepted Ballots Not Accepted Ballots Tabulated Canvass
235,421393 179,161394 54,145395 179,013396 179,013
Figure 8-21: Harris County Returned BBM Discrepancies
Harris County’s Final Vote History reflects there were 179,174 voters who received credit for voting by mail. This value is different than the number of BBMs accepted from their Mail Ballot List and the number of BBMs tabulated.
Tarrant County
Tarrant County provided two reports that addressed ballot by mail statistics. The following chart shows the source and numbers reflected in the source provided by Tarrant County.
393 1120 By Mail Ballot List – Public. This number includes ballots that were returned undeliverable or cancelled by a voter. 394 1120 By Mail Ballot List – Public. This number includes ballots with an OK disposition code. 395 This figure excludes ballots with an OK disposition code and ballots that were returned undeliverable. 396 Harris County Tally Audit Log. This includes limited absentee ballots as well.
Source Ballots Returned Accepted
1120 Absentee List397 83,722 70,479
Ballots Not Accepted
12,905398
Tarrant 2020 Nov ABBM Roster 71,023 70,608
414399
Figure 8-22: Tarrant County Returned BBM Discrepancies
Tarrant County provided a file that reflected there were 70,604 voters who received credit for voting by mail.400 Tarrant County’s Final Vote History reflects there were 66,485 voters who received credit for voting by mail. FAD also reviewed the tabulation audit log that contains a record of all ballots ultimately tabulated. The audit log reflects 65,995 ballots by mail were tabulated. Tarrant County’s canvass report states there were 65,995 ballots by mail cast in 2020.
Requests to Cancel Ballots by Mail
In 2020, there was an influx of voters seeking to vote by mail. In that election, there were also many requests to cancel ballots by mail after they had been issued.
In each of the four counties, the following data was available regarding the number of cancelled ballots by mail:
Collin County
Requested to Cancel BBM and Voted in Person Requested to Cancel their BBM and Voted Provisionally
4,848 voters401 2,382 voters402
397 1120 Absentee List – all return codes. 398 This excludes ballots with an OK disposition code and ballots that were returned undeliverable. There were 338 ballots returned undeliverable. 399 This figure excludes ballots with an OK disposition code. There were no ballots coded undeliverable in this source. 400 1120_Voter History_Exp_SOS_Absentee. 401 Daily Polling Place Mail Ballot Flagging Reports. 402 ABBM Cancelled List – Provisional.
Dallas County
Harris County
Tarrant County
Requested to Cancel BBM and Voted in Person Requested to Cancel their BBM and Voted Provisionally
14,091 voters403 375 voters404
43,572 voters405
11,737 voters407 8,065 voters406
Figure 8-23: Request to Cancel BBM by County
Requests to Cancel Sampled Records Review
A review of a sample of records related to requests to cancel BBMs was conducted to determine whether the counties also properly recorded the voter’s vote history and reported the correct vote history to TEAM.
Collin County
Collin County provided scanned copies of all the request to cancel forms filled out at the polling locations.
Of the 100 records reviewed:
• 96 voters filled out the Request to Cancel form and voted in person.408
o 4 of these voters did not have their ballot and are reflected in Collin
County’s Final Vote History as voting provisionally.
o 3 of these voters appear to have voted during early voting, however,
TEAM reflects they voted on Election Day.
403 Dallas County Ballots Cancelled by Voter. 404 Provisional Ballots – 04-30-2022-12-45-13-PM. 405 1120 Ballot by Mail List – Public. 406 1120 Ballot by Mail List – Public. 407 Absentee Counts by Return Code. There were an additional 126 ballots that were coded as simply being cancelled by the voter. 408 This number is reflected by the statewide database. Of these 94 voters, there were 9 who did not show up on Collin County’s Final Vote History, however, they do appear in the statewide database as having vote history.
• 1 voter filled out the Request to Cancel form and appears in Collin County’s
Final Vote History as having voted early in person. TEAM, however, reflects that this voter returned two absentee ballots that were accepted.
• 1 voter filled out the Request to Cancel form and appears in Collin County’s
Final Vote history as having voted early in person. TEAM reflects this voter’s absentee ballot was accepted.
• 2 voters filled out the Request to Cancel form and never voted.
Dallas County
FAD obtained scans of Request to Cancel forms and performed an analysis on a sample of those records.
Of the 116 records reviewed:
• 42 voters filled out Request to Cancel form and voted in person,409
• 45 voters filled out Request to Cancel form and never voted, and
• 29 voters filled out Request to Cancel form but have vote history for voting by mail in TEAM.410
Requests to Cancel Transmittal Forms
FAD located the transmittal form for requests to cancel ballots by mail between the polling location and EVBB. Documents for at least 51411 of the 61 early voting locations were located. In those documents, 3,214 ballots were requested to be cancelled and surrendered. The polling location with the largest number of requests to cancel was Our Redeemer Lutheran Church, with a total of 318 requests to cancel. Notably, on one of the transfer documents, one of the voters surrendered three BBMs. When properly filled out, these forms would be considered a best practice as they include the number of ballots surrendered, seal numbers documenting the sealed transfer of the ballots, and signatures indicating the seals are verified.412
409 This number is reflected by the statewide database. Of these 42 voters, there were 19 who did not appear in Dallas Final Vote History, however, they do appear in the statewide database as having vote history. 410 This number is reflected by the statewide database. Of these 29 voters, there 17 who did not appear in Dallas Final Vote History, however, they do appear in the statewide database as having vote history. 411 Documents related to an additional location were located however those documents were incomplete and the name of the polling location to which they belonged was unclear. 412 Not all of these forms were properly filled out, as some were missing seal numbers, signatures, etc.
Figure 8-24: Early Voting Request to Cancel BBM
Harris County
Harris County provided a sample of requests to cancel BBMs. Of the 31 voters’ records provided:
• 7 voters filled out a request to cancel form but did not surrender their ballots.
o Of these 7 voters:
▪ All 7 are recorded as having voted early in person on Harris
County’s Final Vote History, even though the forms they filled out indicated they did not vote early.413
▪ 4 of these voters do not have vote history.
▪ 3 of these voters have vote history for voting on Election Day.
• 2 voters returned their ballots but there was no request to cancel form included.
413 One voter’s form is dated 11/11/2020, after the election.
o Both of these voters are recorded as having voted early in person on
Harris County’s Final Vote History, though they both have history for voting on Election Day in TEAM.
• 2 voters filled out Request to Cancel forms but did not properly fill out the form indicating whether or not they surrendered their ballot.
o Both of these voters are recorded as having voted early in person on
Harris County’s Final Vote History, though the on the forms they filled out indicated they voted on Election Day.
o One of these voters did not have vote history in TEAM.
• 20 voters filled out a request to cancel form and voted in person.
o 18 of these voters are recorded as having voted early in person on Harris
County’s Final Vote History, though the on the forms they filled out indicated they voted on Election Day.414
Tarrant County
Tarrant County provided scans of cancelled and surrendered BBMs. Of the 60 records reviewed, 59 voters that cancelled their BBM voted early in person415, and 1 voter cancelled their BBM and never voted.
Other Findings
Dallas County
While on-site in Dallas County, a box of ballots labeled “Cancelled and Cleared” was located. The box contained unopened carrier envelopes for BBMs. Inside this box, several irregularities with regard to BBMs were discovered.
1. Records reflected voter had received credit for voting by mail even though the carrier envelope remained sealed and unopened.
414 One of these voter’s forms did not include the date. 415 This number is reflected by the statewide database. None of these 59 voters appear in Tarrant County’s Final Vote History, however, they do appear in the statewide database as having vote history.
Carrier envelopes in Dallas County were labeled and included an Envelope Tracking ID number as shown in the top right corner of the label:
Figure 8-25: Envelope Tracking ID number on Carrier Envelope
This Envelope Tracking ID number was associated with a particular ballot that had been issued to a particular voter. The number is part of information that is shared between offline counties and the TEAM system. When an event occurs related to that particular ballot such as the issuance of the ballot, receipt of the ballot, acceptance of the ballot, or rejection of the ballot, updates regarding the status of the ballot are provided by the county to TEAM.
Upon review and analysis of the contents of this box, FAD discovered that for 21 of the sealed carrier envelopes returned by voters to Dallas County, Dallas County’s Final Vote History reflected that the voter’s ballot had been counted. Additionally, TEAM reflected the voter’s mail ballot had been accepted. The carrier envelopes that would contain said ballots, however, remained sealed.
2. Records reflected a voter’s timely returned BBM may have been improperly excluded from being processed for tabulation.
One of the sealed carrier envelopes that had been returned to Dallas County contained a notation that the voter had already voted:
Figure 8-26: Note that Voter Already Voted Written on Carrier Envelope
A review of Dallas County’s Final Vote History reflected the individual had not voted in Dallas County. A review of the TEAM database reflected the individual had not voted in Texas in the 2020 General Election.
3. Records reflected a returned sealed carrier envelope for a voter who had not been registered to vote.
One carrier envelope was located that had a notation on a yellow post-it note that read “NR Status.”
Figure 8-27: Not Registered Noted on a Carrier Envelope
A search of this voter’s information in TEAM reflects that the voter’s application to register to vote in Dallas County had been rejected. Dallas County confirmed this voter was not registered to vote in Dallas County in 2020 and never should have been sent a BBM in the first place.
4. Records reflected a returned sealed carrier envelope was not processed apparently due to an error on the label.
One of the sealed carrier envelopes in this box contained notations that the ballot was “canceled” and was “previously returned CL.” CL stands for “cleared.” The unique identifier on the label on this carrier envelope, however, did not match the name printed on the label. Instead, the unique identifier belonged to another individual who had in fact returned his BBM and voted in person. The sealed carrier envelope that FAD discovered appeared to belong to the voter whose name was printed on the label but the carrier envelope was never processed due to confusion between the voter who had cancelled and cleared his ballot and the voter whose name was on the label and was attempting to vote by mail. The voter whose name was on the label did not appear in Dallas County’s Final Vote History, nor did the voter have credit for voting in TEAM.
Figure 8-28: Previously Returned Noted on Carrier Envelope
5. Records reflected that a timely, hand-delivered,416 sealed carrier envelope was not processed.
One of the carrier envelopes located in the box contained a notation that it had been hand-delivered on October 13, 2022 at 11:08 AM.
Figure 8-29: Note that Carrier Envelope was Hand Delivered
The roster for hand-delivered ballots by mail likewise reflects this information.
416 Generally, a marked ballot must be returned via mail or common carrier. Tex. Elec. Code § 86.006 (a). The voter may deliver a marked ballot in person to the early voting clerk’s office only while the polls are open on election day. A voter who delivers a marked ballot in person must present an acceptable form of identification. Id. at (1-a). During the pandemic, the Texas Supreme Court upheld Governor Abbott’s decision to designate a single drop-off location (early voting clerk’s office) per county for those seeking to hand deliver their mail ballots. Under the ruling and the Governor’s October 2020 proclamation, voters had forty days to hand deliver their marked ballot (as opposed to the previous one day—election day—on which this was permitted). See Abbott v. Anti-Defamation League Austin, Southwest, & Texoma Regions, 610 S.W.3d 911 (Tex. 2020). Voters must present an acceptable form of ID, they may only deliver their own carrier envelope, they must sign a signature roster, and then deposit their mail in ballot into a ballot box.
Figure 8-30: Headers on form used by Dallas County in 2020 General Election
Figure 8-31: Date and Time of Delivery and Acceptable Form of ID as filled out in roster for this particular voter
The ballot, however, did not count. This voter does not appear in Dallas County’s Final Vote History as receiving credit for voting. Dallas County’s BBM reports417 do not reflect this ballot was ever returned to Dallas County. Likewise, the TEAM database—which depends on data uploads from the county—does not reflect this ballot was received or accepted though it was hand-delivered to Dallas County.
Harris County
Federal Write-In Absentee Ballots
FAD discovered Federal Write-In Absentee Ballots (FWAB) that were not counted because they were not timely forwarded by San Diego, CA elections officials. These ballots were accompanied by a letter that read:
Dear Elections Official,
Enclosed are FWABs received by San Diego County Registrar of Voters. Theses ballots were received on-time for the November 3, 2020 Presidential General Election.
Number of ballots enclosed (8).
The eight ballots enclosed appeared to have been received by San Diego County prior to November 3, 2020. But the letter was dated November 23, 2020. The package was stamped received by Harris County on November 30, 2020.
417 1120 All ABBM Report for 2020 General Election; 1120 Summary Report for 2020 General Election; List of Voters Sent Mail Ballot.
The FWAB may be used by FPCA voters that have not received their official ballot in order to ensure their vote is received. An FPCA is considered “submitted to the early voting clerk” on the date it is placed and properly addressed in the United States mail.418 Records provided by and related to Harris County FPCA voters do not reflect that Harris County had received the initial FPCA that would provide eligibility for these voters to have voted using a FWAB. The delay in transmittal between San Diego County, California and Harris County, Texas is concerning given the ballots were not mailed until after the local canvass had been completed in 2020, however, it does not appear these ballots would have been eligible for counting due to not meeting the requirement regarding the initial FPCA.
Vote by Mail Review Worksheet
FAD reviewed records – “VBM Review Worksheets” – generated and maintained by Harris County’s SVC and EVBB related to ballots by mail that required additional contact with a voter.
Like any other domain of election procedures, properly filling out and returning a BBM can be subject to human error. In these documents, the issue addressed was primarily in the form of mismatched or missing signatures. When the situation arose, the SVC and the EVBB attempted to address BBM issues with voters so that their ballots could be counted. There were hundreds of completed VBM Review Worksheets that speak to the meticulous nature of the SVC and EVBB correction processes.
The VBM Review Worksheet was used for the following reasons:
• Voter signature did not match the BBM application and the BBM ballot;
• A signature was missing;
• A signed ballot required verification from the voter;
• Missing/incomplete Statement of Residence form;
• No phone number was included in order for SVC/EVBB to contact the voter;
• SVC or EVBB contacted or attempted to contact the voter to correct their signature; and
• A spouse signed for a voter and the SVC/EVBB was unable to confirm because the voter did not answer letters or phone calls.
418 Id. at (i).
The following form shows that two people in a household swapped signatures on their ballots.
Figure 8-32:VBM Worksheet with two people Swapping signatures in one household
This following example shows a voter who responded to the County Clerk’s office calls and sorted out their SOR paper work. After two attempts at contacting them, the voter was able to complete the missing forms online after a reminder from the SVC. The issue was resolved and the voter’s ballot was accepted, as confirmed by the county’s final vote history reflecting this voter had received credit for voting by mail.
Figure 8-33: Resolved VBM Worksheet
Oftentimes these worksheets are marked unresolved because the voter does not respond to the attempted outreach. There were, however, voters whose worksheets were marked unresolved yet they received credit for voting in Harris County’s final vote history record. In other words, the issues outlined in the worksheet appear to have been unresolved despite attempts to fix them but the ballot was counted anyway.
For example, the name on the ballot envelope for the following voter is different from the absentee request signature (presumed to be a family member). It was later determined that the spouse signed both ballots and that both ballots needed to be examined. A signature was unable to be confirmed as a match. Despite multiple phone calls, the issue went unresolved but the voter appears in Harris County’s vote history as having credit for voting by mail.
Figure 8-34: Unresolved VBM Worksheet despite attempts to correct
Figure 8-35: Unresolved VBM Worksheet despite attempts to correct
The following voter was required to provide a statement of residence with their ballot. The SVC attempted to contact the voter but the voter did not leave a phone number. The voter’s registration certification had been returned undeliverable despite the voter’s address being correct and matching the registration. The VBM worksheet is twice marked as unresolved. Harris County’s final vote history reflects that the voter received credit for voting by mail.
Figure 8-36: Unresolved VBM Worksheet despite attempts to correct with voter receiving credit for voting by mail
Another voter did not answer any available phone numbers and did not have an answering machine. The SVC reached out twice but the voter did not answer and the worksheet was marked unresolved. Harris County’s final vote history reflects this voter received credit for voting by mail.
Figure 8-37: Unresolved VBM Worksheet despite attempts to correct with voter receiving credit for voting by mail
The following voter did not complete his statement of residence. The Harris County Clerk’s office attempted to hand him a Statement of Residence Reminder but the voter was marked as “not home.” As a result, their sheet is marked as unresolved, yet this voter has credit for voting by mail in the county’s final vote history.
Figure 8-38: Unresolved VBM Worksheet and incomplete statement of residence
The foregoing examples demonstrate the county’s extensive efforts to resolve issues related to ballots by mail. Efforts involved multiple and repeated attempts to contact the voter using various methods of communication so the voter had a chance to correct any information or provide necessary documentation.
While Harris County’s VBM worksheets are a useful tool for recordkeeping regarding the efforts made by the SVC and EVBB, additional records are necessary to explain why voters with “unresolved” worksheets ultimately received credit for voting.
Records related to the batching and transfer of ballots between the SVC/EVBB and Central Count were located. These records were voluminous and kept in multiple boxes or cases. A sample of the records was scanned to document how the transfer of ballots occurred and the documentation Harris County kept regarding such transfers. These forms and procedures would be considered a best practice in documenting the transfer of ballots between the involved entities.
As noted above, Harris County organized, stored, and transferred ballots by mail using numbered tubs. A document with the batch number of the ballots scanned and an accounting of the total ballots included was maintained. Batch cover sheets contained the batch number and a list of the voters whose ballots were being scanned. Records in Harris County included a cover sheet denoting how many ballots had been accepted by the EVBB, a breakdown of the ballots to be scanned, ballots that required duplication, empty carrier envelopes, and carrier envelopes containing more than one ballot. The transfer of ballots properly involved the use of seals and Harris County maintained the seals, a sheet documenting how many ballots were being transferred for scanning, and a sheet with the seal numbers included.419
419 Harris County used forms for the transfer of BBMs between the EV Clerk and EVBB that were virtually identical in format as the forms used for the transfer of provisional ballots between the EV Clerk and the EVBB. These forms, if properly filled out, are a best practice.
Figure 8-39: Harris County EVBB Batch Cover Sheet
Figure 8-40: Batch Tub Cover Sheets
